ANOXIA TEST FOR MYOCARDIAL ISCHAEMIA · THE ANOXIA TEST FOR MYOCARDIAL ISCHAEMIA BY NORMANCOULSHED FromtheLiverpoolRegional Cardiac Centre, Sefton GeneralHospital, Liverpool ReceivedMay11,

THE ANOXIA TEST FOR MYOCARDIAL ISCHAEMIA

BY

NORMAN COULSHEDFrom the Liverpool Regional Cardiac Centre, Sefton General Hospital, Liverpool

Received May 11, 1959

When the symptoms are typical, the diagnosis of angina pectoris is relatively easy, even in theabsence of objective signs of cardiovascular disease. Nevertheless, Evans (1952) pointed out thatwhere this objective evidence is missing the diagnosis of angina pectoris is likely to be wrong in oneout of every four cases. Biorck (1946) stated that he could be certain of the diagnosis in onlyone out of every six patients who had been thought to have angina pectoris.

To help in difficult cases, where objective evidence of coronary insufficiency cannot be foundor relied upon, four types of special tests have been designed to produce electrocardiographicevidence of coronary insufficiency. Levine et al. (1930) introduced the adrenaline test, and Ruskin(1947) the pitressin test, but both have been abandoned for general use, as they are potentiallyharmful. The exercise test has been widely used and developed (Master et al., 1935 and subse-quently, Evans and Bourne (1941), Twiss and Sokolow (1942), Biorck (1946), Wood et al. (1950)and many others). Unfortunately, there has been no general agreement as to what constitutespositive evidence of coronary insufficiency in this test, and this makes assessment of its value alittle difficult.

The fourth method by which evidence of coronary insufficiency can be demonstrated is by theproduction of general anoxia by inhalation of a low-oxygen gas mixture, and recording electro-cardiograms at certain intervals. Extensively investigated by Levy and his co-workers (1938,1939, etc.) by Patterson et al. (1942), and Mathers and Levy (1950, 1952), the test produced a posi-tive result in only half of the patients with known coronary insufficiency. However, Levy's threecriteria of positiveness in the test have been universally used and have formed the basis of allsubsequent work in this field: they were as follows.

1. The arithmetical sum of the RS-T segment deviation in the three standard leads and theprxecordial lead totals 3 mm. or more, when 1 cm. deflection is equivalent to 1 millivolt.

2. There is partial or complete reversal of the T wave, with RS-T segment deviation of 1 mm.or more, in lead I.

3. There is complete reversal of the T wave in the prwcordial lead.Malmstrom (1947) modified the method of performing the test by reducing the concentration of

oxygen in the inhaled gas mixture and added carbon dioxide, successfully increasing the number ofpositive results in patients with coronary insufficiency, without producing abnormal records inpatients where this insufficiency did not exist. His series of patients was small in number, however.This paper is concerned with the findings and personal observations in a larger series of patientsupon whom Malmstrom's modification of Levy's anoxia test was performed.

METHOD OF APPLICATIONTo induce anoxia, Levy and his associates used a mixture of 10 per cent oxygen and 90 per cent

nitrogen, but for reasons that will be discussed later, Malmstrom employed a mixture of 6 5 per79

on Novem

ber 28, 2020 by guest. Protected by copyright.

http://heart.bmj.com

/B

r Heart J: first published as 10.1136/hrt.22.1.79 on 1 January 1960. D

ownloaded from

http://heart.bmj.com/

cent oxygen, 4-5 per cent carbon dioxide, and 89 per cent nitrogen. A similar mixture was usedthroughout this series, using a standard Walton anxsthetic machine.

The mixture of gas is delivered to the patient through a system of flutter valves to prevent re-

breathing, and an emergency supply of oxygen is ready to be introduced into the circuit if necessary.The gas can be inhaled by the patient either through a standard anesthetic face piece or through a

mouthpiece, using a nose clip to prevent nasal breathing.All patients in this series were seen and examined by the author, before the test was performed.

The object of the test and its method were fully explained to each patient, and any undue apprehen-sion or reluctance to undergo the test was sufficient reason for it not to be used. This was not onlyfor the sake of safety, but because it has been shown that fright in itself is sufficient toproduce electrocardiographic changes (Mainzer and Krause, 1940). No vasodilator drugswere given for twelve hours before the test, and only one case was being treated with digitalis.The test was performed at least two hours after the last meal, and after the patient had been at restfor at least half an hour. Instructions were given that any discomfort experienced during the testother than moderate dyspneea, should be signalled by raising an arm, when the test would be stopped.For convenience, all the tests were performed with the patient semi-recumbent on a couch, no

alteration of the position being allowed. All the electrocardiograms were recorded on a direct-writing instrument, the tracings being thus available for immediate inspection.

A full twelve-lead electrocardiogram was recorded before the test was begun, the chest lead tobe chosen for use during the test being that with the tallest R and the smallest S, this lead beingassumed to be over the maximum thickness of the left ventricle. With all the preliminaries comple-ted, the mouth piece or face piece was applied and the patient allowed to breathe room air for a

few minutes to become accustomed to the apparatus. The gas mixture was then introduced, a

constant flow being kept up during the test. Records were originally taken at fives ten, fifteen, andtwenty minutes anoxia, but it was soon realized that prolongation of the test beyond ten minuteswas unnecessary, positive results often occurring within the first few minutes. Finally, therefore,the technique evolved was to record electrocardiograms at two, five, seven, and ten minutes, thenafter the inhalation of oxygen for one minute, and again after five minutes rest, repeating therecording at five minute intervals until the tracing returned to the resting state. If the patient sig-nalled the presence of discomfort or if the records showed notable change, the anoxia was stopped,an electrocardiogram taken, oxygen given for one minute, and further recordings made as describedabove. The three standard leads of the electrocardiogram, and one chest lead, were recorded ateach interval, and with practice it was found possible to complete each set of tracings well withinthirty seconds. No patient had the test performed if there was any suspicion of myocardial in-farction having occurred within the previous four months. Other conditions precluding the testwere congestive cardiac failure, obvious illness of the patient, severe anamia, and a high degree ofpulmonary disease. The exercise test was done according to the method of Wood et al. (1950).

The interpretation of the electrocardiograms was based upon Levy's work, and it was confirmedthat RS-T segment changes were the most important findings. Compared with the control group,

no significant alterations occurred in the heart rate, the P waves, and the QRS complexes, andsignificant arrhythmias occurred in only two patients. The measurement of RS-T segment deviationwas made in each lead by taking the average of five successive complexes. The iso-electric level wastaken as the P-Q interval, or the point before the first deflection of the QRS complex, as recommen-

ded by Levy (above), Malmstrom (1947), Katz (1946), and Master (1953). The T-P level wasimpractical, since the tachycardia found in many of the tests obliterated this interval. The RS-Tsegment level was taken as the point when this interval began. The arithmetical sum of RS-Tsegment deviation in all four leads could then be found.

RESULTSOne hundred and eleven patients were examined, 126 anoxia tests being performed. Eleven

exercise tests were performed for comparison with the anoxia test. The follow-up period varied

NORMAN COULSHED80

on Novem



/B


ownloaded from


THE ANOXIA TEST FOR MYOCARDIAL ISCHJ4EMIA

from eighteen months to three years, during which time six of the patients died, autopsies beingperformed on three.To simplify observation and assessment, the patients were divided into four groups.Group I was made up of 41 patients with no clinical evidence of organic cardiovascular

disease, and served as a control.Group 2 included 53 patients in whom coronary artery insufficiency was suspected on clinical

grounds.Group 3 comprised 7 patients with peripheral arterial disease causing symptoms of intermittent

claudication, but without symptoms of angina pectoris.Group 4 was a miscellaneous group of 10 patients.The distribution of cases in Group 1 and in Groups 2, 3, and 4 combined, as regards age and

sex, was approximately equal.Twenty-five tests on twenty-one patients were abandoned before the test could be completed.

The main reasons for this failure were fear of the test in twelve cases, and intolerable dyspncea ineleven. Fear amounting to panic was probably the result of inadequate explanation of what washappening, and rarely occurred as experience in the use of the test by the operator increased. Allthe patients in whom dyspncea caused the test to be stopped had clinically recognizable pulmonaryemphysema.

Constricting chest pain exactly similar to anginal pain occurring on effort was present in 17of the tests. Electrocardiograms taken at the time of the pain showed positive results in 16.The only case where no significant cardiographic change occurred had previously shown apositive test. On the other hand, the test was stopped on 15 occasions because of strikingalteration in the electrocardiogram without the occurrence of any discomfort whatsoever, one ofthese patients having a total RS-T segment deviation of Il 0 mm. This experience is in line withthat of other observers, in that pain and cardiographic change are not always concurrent; it isuncommon to have pain without cardiographic change, but not vice versa. Twenty-three patientshad the tests stopped because of such changes but admitted later to having very minor degrees ofchest discomfort. When the pain complained of was not typically cardiac, as happened in threeinstances, there was no cardiographic change. In all the other cases pain was immediately relievedby the administration of oxygen, although the return of the cardiogram to the resting state was in-variably much less rapid, in some cases taking up to fifteen minutes.

Table I lists the types of reaction to the test observed in this series, although not all these reactionscaused the test to be abandoned. Headache was described as bursting, but was relieved at once by

TABLE IREACTIONS TO THE ANOXIA TEST

Number of times Number ofobserved patients

Fear .. .. .. .. .. 19 16Intense dyspncea .. .. .. 21 20Chest pain, cardiac type .. .. 18 16Chest pain, non-cardiac type .. 5 5Headache .. .. .. .. .. 3 3Blurred vision .. .. .. 1 1

oxygen, as was the one example of blurred vision. None of the reactions to the test produced anypermanent ill effects.

Table II summarizes the results in the whole series of completed tests. For purposes of moredetailed analysis, group 2, clearly the most important, has been broken down into further subgroups,A, B, C, and D. Group A included ten patients where coronary artery disease was known to exist,G

8I

on Novem



/B


ownloaded from


NORMAN COULSHED

TABLE IXSUMMARY OF THE RESULTS OF 102 ANOXIA TESTS

Control Positive resultselectro- Total number

cardiogram of tests Number Percentage Combinedpercentage

Group 1 Normal 35 0 0 0

Group 2 Normal 36 29 81 86Abnormal 20 19 95

Group 3 Normal 5 3 60 60Abnormal 0

Group 4 Normal 5 0 0Abnormal 1 1

because of previous myocardial infarction. Eight tests were completed, and six were positive.No explanation can be offered for the negative test in one case, but the other patient was not sufferingfrom anginal pain at the time of the test, and was in fact performing very heavy manual labour, somuch so that he cannot have been suffering from coronary artery insufficiency at that time. Hisexercise test also was negative.

Group B was made up of ten patients who had no history to suggest myocardial infarction, butin each case the electrocardiogram at rest was abnormal, suggesting coronary artery insufficiency.Eleven tests were completed in this group, and each was positive (Fig. 1 and 2).

Group C comprised fourteen patients who had normal or only slightly abnormal electrocardio-grams when at rest. In each case the clinical story was suggestive of angina pectoris, but othersources of pain were present, such as cervical spondylosis, hiatus hernia, peptic ulcer, chest wallinjury, etc. Nineteen anoxia tests were completed on thirteen of these patients, and fifteen werepositive. Taken as a whole, however, the test was positive at least once in twelve out of the thirteenpatients (Fig. 3 and 4). The one patient in whom the tests were negative had a very dubious historyof anginal pain and a proven duodenal ulcer; his second test, performed six months after the first,was done when he had suffered no pain for six months, and when he was seen some three yearslater no more discomfort had occurred; it is very likely, therefore, that in this case the test wascorrect and there was, and had not been, any coronary insufficiency.

Group D included nineteen patients who were similar to Group C except that no alternativesource could be found for their chest pains. Eighteen tests were completed, and sixteen were posi-tive (Fig. 5 and 6). One of the two negative tests was done on a patient whose angina had beenrelieved by mitral valvotomy, the pre-operative test being positive. The other test which wasnegative occurred in a patient in whom the history was entirely typical of angina pectoris, and apositive anoxia test had been confidently predicted.

In Group C, the test proved of great value. It established the presence of coronary insufficiencyin twelve out of thirteen patients in whom the disability was thought to exist, and it was negative ontwo occasions when used on a patient where the clinical diagnosis of angina pectoris was almostcertainly wrong.

Group D was also of great interest. The patient with mitral stenosis and angina pectoris had apositive test before valvotomy relieved her symptoms, the anoxia test becoming negative with theimprovement in the clinical state. Only one did not have a positive test, and the others had at leastone positive test establishing the presence of coronary insufficiency. Table III summarizes thefindings in subgroups A, B, C, and D.

Returning to the other groups of patients, Group 3 comprised seven patients with intermittentclaudication, none of whom would admit to any chest pain. Anoxia tests were completed on five

82

on Novem



/B


ownloaded from


THE ANOXIA TEST FOR MYOCARDIAL ISCH4MIA 83

and there was a positive result in three. In addition, three patients in Group 2 had claudicationpains accompanying their angina pectoris, and in each of them the anoxia test was positive. Innone of these cases was it possible to perform exercise tests, and it is in these circumstances that theanoxia test has a great advantage over the exercise test.

wnz9.,Mna

~~~~~Vs

. .': .:

AVWs

~~~~Vs

5et'|'tVs

I 11

OU44, OWV"f. OVf tA'L 6

I . a

esrznz |

t.X..s ... F

i.. ,j. .;....s , 1-it W. wi.

, ,^ .

.. 5 i t, _",

:v.: :. .. ..,.... _, :

*::: ;:. :^' '.

- ::' '^.:^,.+ .tA F

I

iO._ _Rv 4...

i i t;+ A

;;i3 t. ,.'.i 2 fs,.._.. ffi<.o... g L't. z ",D' w-

i.i te*_, _.., ..^'. __ ,.s,.,._ _ .,.+,

TtsxAr n .1..':l

| :. ::_^. ...^.Z:k +. .:. ..s ^ ir3 !_-, , %v

t'. '#}s + ^v>^_t_^+='*.... ^_, t b

FIG. 1.-Women, domestic worker, aet. 45. Complained of leftsubmammary pain for three years. Regarded as neurotic,and given E.C.T. B.P. 155/95, heart size and sounds normal.Anoxia test stopped after one minute because of substernalpain. Test positive, with total RS-T deviation of 6 mm.(criterion 1), and T wave reversal with RS-T deviation ofmore than 1 mm. in lead I (criterion 2). Exercise test pro-duced inversion of TI and TV5 with total RS-T segmentdeviation of 4 mm. This test was positive also.

- '4It.I

on Novem



/B


ownloaded from


84~~~~NORMAN COULSHED

Ri o

3m tVA

I~~~~~~~~~~PE.ffr 7,T

t A~~~~~~~~~~~~~i ~ ~

V4

FIG. 2.-Man, aet. 59, electrician's mate. Substernal pain for twelve months, brought on by exercise,but present also after meals, relieved by alkalies. Exercise test at onset of pain was negative,and barium meal showed duodenal ulcer. B.P. 160/90. Heart size and sounds normal. Anoxiatest stopped after one minute because of substemal pain. Total RS-T segment deviation of2-0 mm. (criterion 2) and reversal of TV4 (criterion 3).

PAh

4I

FIG. 3.-H-ousewife, aet. 62, Epigastric pain for two years, worse after exercise and aftermeals, relieved by alkalies and not by nitrites Heart size and sounds normal, B.P. 140/90.Barium meal showed a duodenal ulcer. Anoxia stopped at five minutes because of cardiographicchange and slight substemal pain. The test was positive, with total RS-T segment deviation of4*5 mm. partial inversion of Tl with RS-T deviation of 1 mm. and complete inversion of TV4.She died in a paroxysm of pain six months later.

84

on Novem



/B


ownloaded from


THE ANOXIA TEST FOR MYOCARDIAL ISCHAfMIA

TABLE IIIRESULTS OF ANOmA TESTS IN 53 PATIENTS WrrH SUSPECTED CORONARY INSUFFICENCY

Group 2 Number of Number of Number of Number of patientsSubgroups completed positive patients with at least one

tests tests positive test

A 8 6 7 5B 11 11 8 8C 19 15 13 12D 18 16 17 16

Totals .. 56 48 45 41

Included in Group 4 were five patients who had symptoms and X-ray confirmation of diaphrag-matic hiatus hernia. In none of these were the symptoms suggestive of coronary insufficiency, andthe anoxia tests completed on three of them were all negative. In contrast to this, four patients inGroup 2 also had radiologically proven hiatus hernia, but symptoms suggestive of true angina pec-toris. The anoxia test was positive in each, thus supporting the clinical impression that hiatus herniaalone does not produce anginal pain. If the defect is present and the patient has angina, he or shehas also coronary insufficiency. The anoxia test was completed on one patient in Group 4 who hadmitral stenosis without angina, and it was negative. Six in Group 2 had mitral valve disease withangina, and the test was positive in each.

Two patients had anoxia tests performed in an effort to elucidate the etiology of fainting attacks.One had attacks that were thought to be non-organic, and his anoxia and exercise tests were negative.The other was suspected of having Stokes-Adams syncope, but he was never observed in an attack,and his resting cardiogram was only slightly abnormal. His anoxia test was strikingly positivealthough he had no chest pain, no clouding of consciousness, and no change in cardiac rhythm(Fig. 7); the diagnosis of coronary insufficiency was confirmed later at a coroner's necropsy.

Of the fifty-two tests that were positive, nineteen became so after two minutes anoxia or less;twenty-seven at between five and ten minutes anoxia, and only one after the anoxia had ceased.Table IV summarizes the relative frequency of the criteria of abnormality, the figures being similarto those of Weintraub and Bishop (1947) and of Levy et al. (1941). These figures confirm the view

TABLE IVSHOWING RELATIVE FREQUENCY OF THE CRITERiA OF ABNoRMALrry IN THE ANOXIA TEST

Number of Percentage ofCriterion tss total positivetests ~~tests

1 Alone .. .. 37 73Alone or combined 49 96

2 Alone .. .. 0 0Alone or combined 7 13-7

3 Alone .. .. 2 3-8Alone or combined 11 21-6

1 and 2 .. .. 3 5-8

1 and 3 .. .. 5 9.7

2 and 3 .. .. 0

1, 2, and 3 .. .. 4 7-7

85

on Novem



/B


ownloaded from


86~~~~NORMAN COULSHED

RSIGRE-COO .

FIVE MIMUTLS AN4OXIA

FIVE MINUITES AMT( OXYGEN4

I 1! 3 ~~~~~V4-Fso. 4.-Housewife, aet. 47. In 1952 angina

pectoris diagnosed on the history, no abnormalphysical signs were present. Four years laterthe pain had been distorted by a gross anxietyoverlay and the presence of a gastric ulcer.Still no abnormal cardiovascular signs, andB.P. normal. Anoxia test was positive, withtotal RS-T segment deviation of 5-5 mm. Nopain occurred.

LISTING RICOiLS

.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...

TWO MINUTES ANIOXIA

I~~~~~~~~~~~~~~~~:! L I..i

FVvE~ MImUTEs AnaI O*saoE4

FIG. 5.-Housewife, aet. 51. Severe epigastricand lower substemal pain coming on

after minimal effort, referred to left arm, left

thigh and knee. B.P. 140/90, heart size and

sounds normal. Anoxia test was positive(criteria and 2), with total RS-T segment

deviation, of 12-5 mm. coinciding with slightsubsternal pain. The patient died six months

after the test, in a sudden paroxysm of pain.No autopsy was done.

of Turner and Morton (1952) that RS-T segment deviation was the most constant and reliablefinding in the anoxia test. Fig. 8 depicts the total RS-T segment deviation in all four leads, in thecontrol series and in the abnormal patients. It shows that the maximum of 2-5 to 2-9 mm. for anormal test is correct, and that 3-O mm. and above is probably correct for those patients withcoronary insufficiency. The maximum RS-T segment change occurred most frequently in theprrcordial lead, being more than three times as often as in the other leads together, in the patientswith positive tests.

The heart rate increased on an average by 26 beats a minute in the control series, and by 32

86

on Novem



/B


ownloaded from


THE ANOXIA TEST FOR MYOCARDIAL ISCH4EMIA 8

FO&* mMWUTEs AEIOKIA

F-iIEI PFTtE" "MIUTIs AriEa Owm.E

Fioi. 6.-Housewife, aet. 46. History of typical angina pectoris for twelve months, with noabnormal physical signs. B.P. 130/80. Anoxia test stopped at 4 minutes because ofsubsternal pain. Total RS-T segment deviation 4-5 mm. (criterion 1), reversal of TI andTV4 after the anoxia was stopped, with a slow recovery rate. RS-T segment deviation of1 mm. in lead 1 with reversed T wave constitutes criterion 2, and the reversal of TV4criterion 3.

.h.ti.......ot

Fwh MrnuTh. AN40ZIA

~~~-4~~~~~~~fl ~~~~~~~~r.....-.

I i_: Z N

0 h4MUIE CO O

F6W MI4UWs AFTIL MICH&r

-r,I

I;

I £ V4

FIG. 7.-Man, aet. 54, motor driver. No chest pain, but attacks of syncope when pulse ratealleged to fall to 40 a minute. Normal heart size and sounds, B.P. 155/90, negative clinicalexamination. Anoxia test stopped at five minutes because of gross cardiographic change.No pain occurred. Note the tracings are half sensitivity, and allowing for this, RS-Tdeviation was equivalent to 13-0 mmn. (criterion 1); TI was reversed with RS-T deviationof more than 1 mm. (criterion 2); TV4 was completely reversed (criterion 3).

.jjjj

87

on Novem



/B


ownloaded from


NORMAN COULSHED

COI4LETED TEsT3 in D(O xL GwoUPS

iI

V9)

I

COMLETEo TEsT rm SU3ECTIDCoONMY INsUFFIcI.cy-

IiI

I IwI

I_

%h 0 qll 3r; "Ih ol 4h b to

R5-T 5EGMENT DtEVIATION [MPS.JFIG. 8.-Total RS-T segment deviation in control series and in patients with coronary insufficiency.

beats a minute in the series under test, but the increase in rate bore no relation to the outcome of thetest. Unless T wave changes amounted to total reversal in lead I or the prxcordial lead, they alsohad no relation to the outcome of the test. No T wave reversal occurred in lead II unless it was pre-sent in lead I also, and when present in lead III it had no relation to the result of the test.

Prolongation of the P-R interval occurred in only one case, alternating ventricular extrasystolesin one, and a wandering pacemaker in a third. These were the only instances of either conductiondefects or arrhythmias in this series, and previous investigations have also found such occurrencesuncommon. Roehm et al. (1952) devised a method of calculating what was referred to as the"anoxnmia index" in this type of test. This depended upon the prolongation of the QT intervalin patients with coronary insufficiency being made greater in the anoxia test. The index wasworked out on all the patients in this series, but the difficulties of measuring the QT interval arewell known, and the results of the calculations in this series showed an equal number of positive andnegative indices in the control patients and in the abnormal group.

Exercise tests were performed on eleven patients for comparison with the anoxia test. In onethe anoxia test was positive when the exercise test was negative-this patient had a myocardialinfarction some twelve months later. The exercise test was never positive when the anoxia test wasnegative; five patients had both tests positive. When pain was provoked in the exercise test, itinvariably lasted longer than with the anoxia test, there being no oxygen ready to relieve thesymptoms.

Since the investigation began, six of the patients are known to have died. None of these wasin the control group, and three who died suddenly were certified as having coronary thromboses, butno autopsies were performed. Autopsies were carried out on the other three. One man of 62died in a paroxysm of pain, autopsy showing a recent septal infarct, with gross coronaryatherosclerosis, and areas of mainly subendocardial fibrosis in the left ventricle. Another diedtwo years after the anoxia test had been done, and was shown at autopsy to have an unsuspected

88

III

III

on Novem



/B


ownloaded from


THE ANOXIA TEST FOR MYOCARDIAL ISCHAMIA

calcific aortic stenosis, much coronary artery thickening, and a recent intramural infarct, butno evidence of subendocardial fibrosis. The third patient died suddenly, some six months afterhaving a positive anoxia test: he had complained of recurrent syncopal attacks, in one ofwhich he was found to have a very slow pulse, but at no time did he complain of any chest pain.He died in one of the attacks, and the necropsy showed widespread severe coronary atherosclerosis,with diffuse myocardial fibrosis, much of it being subendocardial.

ELECTROCARDIOGRAPHIC CHANGES IN THE ANOXIA TESTThe predominant abnormality produced, namely RS-T segment depression, for a long time

has been known to occur also in spontaneous paroxysms of angina, and Parkinson and Bedford(1931), and Wilson and Johnston (1941) pointed out that similar changes of a more permanentnature occurred with coronary occlusion, the underlying cause being presumably of a like naturein each case, but temporary in one and permanent in the other. Brow and Holman (1934)reported that similar cardiographic records had been obtained in patients who had spontaneousattacks of angina, and in animals subjected to progressive anoxia.

Riseman et al. (1940) produced similar records in patients with anginal pain precipitated byexercise or anoxia, and found that the inhalation of oxygen could delay the onset of the changes inthe cardiogram precipitated by exercise. Records during attacks of anginal pain have also beenrecorded by Bousfield (1918), Feil and Siegel (1928), Wood et al. (1931), Hall (1932), Turner (1933),Shapiro and Smyth (1937-8), Bryant and Wood (1947), and Scherf and Boyd (1948), and nearlyall showed predominant RS-T segment depression in the standard leads, and in the prxcordiallead when the latter were recorded.

Similar tracings are shown in Fig. 9, from a woman aged 71. She had typical angina oneffort, but the upper record was taken when she had a short attack of pain, whilst lying at rest. The

1I11 DL VzE4 V6

I EL 111 VI-I I

FIG. 9.-The upper record was taken during a paroxysm of anginal pain, the lower recordtwenty four hours later, when no pain was present.

lower record was taken twenty-four hours later, and although it is still abnormal, the marked RS-Tsegment depression in leads I, V4, and V6 has largely disappeared.

Myocardial anoxia produced by severe anemia can produce records identical with those seen inthe anoxia test. Fig. 10 shows the electrocardiogram of a woman, aged 52, taken half an hour beforeshe died as a result of a massive intraperitoneal haemorrhage. The gross RS-T segment depressionwas unaccompanied by any macroscopic change in the heart, and the coronary arteries were normal.

Master et al. (1941) describing the clinical, electrocardiographic, and autopsy findings in cases

89

on Novem



/B


ownloaded from


90 NORMAN COULSHED

I EI m

.4.

CLA 41VL O.

VI V3 \4 V5 V6

FIG. 10.-Electrocardiogram taken from a patient with severe aneemia, half an hourbefore death.

of acute coronary insufficiency, demonstrated again the presence of RS-T segment depression andcorrelated it with the "diffuse foci of myomalacia, usually mainly subendocardial, in the papillarymuscles of the left ventricle and the interventricular septum." Thomson and Feil (1944), Price andJanes (1943), Bailey (1946), and Pirani and Schlichter (1946) have recorded similar RS-T depressionfound in patients who at autopsy were shown to have subendocardial infarcts; and experimentalproduction of endocardial damage has the same result (Boyd and Scherf, 1940; Wolferth et al.,1945; Pruitt and Valencia, 1948). The theoretical concept of these patterns has been expoundedby Bayley (1943, 1944, and 1946).

It is probable therefore that the RS-T segment depression occurring in spontaneous angina,anmmia, and the exercise and anoxia tests is due to subendocardial muscle damage. Biorck(1948-9) suggested that this damage occurs because the subendocardial muscle is the last to receiveits oxygen supply from the penetrating branches of the coronary arteries, and is probably sensitiveto oxygen lack.

T wave changes alone have been relatively uncommon in all the published series of anoxia tests,but in this present series, they have occurred in the patients who appeared more severely affected,and when present, they have persisted longer than RS-T segment changes. It is likely that theyrepresent more advanced coronary insufficiency but there is no proof of this.

DISCUSSIONGreene and Gilbert (1921 and 1922) demonstrated that sufficient anoxia will produce cardiographic

abnormalities even in healthy subjects, but the different types of anoxia tests have been designed to producecharacteristic and recordable changes in as many pathological cases as possible, while causing no change ofsignificance in normal subjects. Dietrich and Schwiegk (1933), using a mixture containing about 8 per centoxygen in nitrogen, were able to produce abnormal cardiographic patterns in 35 out of 45 patients with heartdisease. Rothschild and Kissin (1933) found a mixture of 6 per cent oxygen in nitrogen caused their patients

on Novem



/B


ownloaded from


THE ANOXIA TEST FOR MYOCARDIAL ISCHk4MIAto lose consciousness, and reverted to an unsatisfactory rebreathing technique, as did Katz et al. (1934),whereby the test could never be properly standardized. Larsen (1936) used a 9 per cent oxygen in nitrogenmixture, and produced 13 positive tests among 30 subjects with coronary disease. The same mixture hasbeen used by Mannheimer et al. (1946, 1948) who have found the test useful in the assessment of certaintypes of congenital heart disease.

Levy and his associates have performed many hundreds of anoxia tests using a mixture of 10 per centoxygen in nitrogen, the patient breathing the gas for a maximum of twenty minutes. They found that thepresence of a positive test bore no relationship to the levels of blood oxygen at the time, these varying from67 to 83 per cent saturation; this work has been confirmed by Mathers and Levy (1950). However, Levycould only produce 55 per cent positive tests in subjects with coronary disease, and suggested that the pro-duction of pain without cardiographic abnormality should be regarded as a presumptive positive test.Similar figures were produced by Patterson et al. (1942) and Pruitt et al. (1945) but no false positive tests werefound. Weintraub and Bishop (1947) alleged that reactions denoting ischlemia occurred in nine out oftwo hundred healthy people, but in fact none of these nine could be regarded as normal controls.

Penneys and Thomas (1948, 1950) attempted to modify the test by producing a more constant andstandard lowering of the arterial oxygen level, but it made the test very difficult to perform, and did nothingto modify the established criteria of abnormality. Hecht (1949) used a mixture of 10 per cent oxygen,3 per cent carbon dioxide, and 87 per cent nitrogen, and was able to produce a positive test in 74 per centof his patients with angina pectoris, all the controls being negative. Using Levy's technique, Alexander et al.(1950) could produce only 33 per cent positive tests in subjects with coronary disease.

Nylin (1943), Biorck (1946 a & b), Lindgren (1946 a, b, & c) Soupanki (1948) were Scandinavian investi-gators whose results were essentially similar to those of Levy and his associates. Malmstrom (1947) con-sidered the method giving 10 per cent oxygen in nitrogen produced not only a wide variation in blood oxygenlevels at the end of the test, but also a very variable rate of fall of the oxygen level, both making standardiza-tion of the test more difficult. In addition to this, hyperventilation alkalosis could be produced, which itselfcould cause RS-T segment depression.

In order to produce a more rapid fall in the blood oxygen level, and to standardize the test better, Malm-strom used a mixture containing 6 5 per cent oxygen, and added 4-5 per cent carbon dioxide, to cause hyper-ventilation and a rapid lowering of blood oxygen, without alkalosis. He found that alveolar oxygen tensionquickly reached a constant low level, there was less variation in the respiratory response, and no alterationin blood pH. No abnormal results were found in his control subjects, and eight out of thirteen patientswith myocardial ischmmia had positive tests.

Master (1953) has admitted that the exercise test produces 6 per cent of positive results in patients withoutmyocardial ischemia, although he claims that it is positive in 95 per cent of coronary insufficiency. Never-theless, the finding that one out of every 20 patients without this condition has a positive test detracts greatlyfrom its value. This does not occur with the anoxia test, and none of the 35 patients in the present controlseries had a positive test. Moreover, with positive tests in over 90 per cent of pathological cases, Malmstrom'smodification of the anoxia test has removed the greatest objection to its use, namely that only a relativelysmall percentage of patients with coronary insufficiency could be demonstrated to have this condition.It has other advantages: it can be used on patients who are unable to undertake physical exercise, there isconstant electrocardiographic control, and oxygen is immediately available making the procedure definitelyless dangerous. It must be admitted, however, that it is more time-consuming and that it cannot be used onpatients with a low respiratory reserve.Comparing the two types of test, Master et al. (1944) were able to produce similar records with each test

on a series of cases of myocardial ischiemia. Biorck (1946) found 27 per cent of patients had positive anoxiaand negative exercise tests, 18 per cent had positive exercise and negative anoxia tests, and 55 per cent hadboth positive. In the present series, only a small number of exercise tests were performed, and in only onecase was the anoxia test positive and the exercise test negative; there were no patients with positive exerciseand negative anoxia tests. It is probable that a place exists for the use of both tests, but the anoxia test is morereliable and less dangerous and should be the method of choice.

It has been proved of value when intermittent claudication prevented the performance of an exercise test.Three out of five cases showed evidence of coronary insufficiency and this type of patient had to be treatedcautiously. Used in an attempt to elucidate the nature of chest pain in a patient with a hiatus hernia, it hasbeen shown that the latter alone is insufficient to produce a positive test, which appears only when the,pain is due to myocardial ischamia, the hiatus hernia being incidental.

91

on Novem



/B


ownloaded from


NORMAN COULSHED

SUMMARYThe modification of Levy's anoxia test as suggested by Malmstrom has been used to investigate

a series of patients. It has been found satisfactory in that it increases the number of positiveresults in patients with coronary insufficiency to a level comparable with that obtained in theexercise test, without producing any false positive results.

Thirty-five patients with no clinical evidence of organic cardiovascular disease served as a controlgroup, and in none was a positive anoxia test produced.

Where coronary insufficiency was known to be present or strongly suspected on clinical grounds,56 tests were completed and 48 were positive (86%). This latter figure is misleading in that onetest was performed on a patient with known coronary disease but with no symptoms of coronaryinsufficiency; two tests were negative and followed by a clinical course suggesting that the painwas not cardiac in origin; and one test was negative in a patient whose symptoms had been relievedby mitral valvotomy. Discounting these four tests, 48 tests were positive out of 52, or 93 per cent.Analysis showed that 41 cases out of 43 were found by the test to have coronary insufficiency incircumstances in which it was expected that the test would be positive, giving a total percentage of95.

The test was positive in 3 out of 5 patients with intermittent claudication but no symptoms ofcoronary insufficiency, and was strikingly positive in one patient suspected of having Adams-Stokessyncope. It was negative in patients with diaphragmatic hiatus hernia when coincident coronarydisease was not suspected.

The anoxia test is more time-consuming than the exercise test, but is simple to perform andprovided reasonable care is taken, it is not dangerous; it has the advantage over the exercise test inthat the patient is under continuous electrocardiographic observation, a supply of oxygen isconstantly available so that the test can be stopped at once and its effects reversed very quickly. Itcan be used on patients who are physically incapable of exertion, and because of the absence ofmuscle tremor, the electrocardiographic records are better than those in the exercise test.

When graphic proof of coronary insufficiency is being sought, it is better to use the anoxia testfirst, but it is probable that there is a place for both types of test. Neither should be used as aroutine investigation, as these methods cannot replace careful clinical assessment.

The major part of this work was undertaken as part of an M.D. thesis in the University of Liverpool, and I amgrateful to the Professor of Medicine, Lord Cohen of Birkenhead, for granting me permission to publish the work.

I would also like to thank Drs. E. N. Chamberlain, E. Wyn Jones, W. S. Sutton, and C. S. McKendrick for allowingme access to their patients in the Regional Cardiac Centre, Sefton General Hospital, Liverpool, and the LiverpoolRoyal Infirmary. In both hospitals I have had cheerful and willing co-operation from the technicians in the electro-cardiographic departments, and without their help this investigation could not have been undertaken.

REFERENCESAlexander, S. F., Wittenberg, S. J., Brown, E. T., and Koffier, A. (1950). N. Y. State J. Med., 50, 535.Bayley, R. H. (1943). Amer. Heart J., 26, 769.

(1944). Amer. Heart J., 27, 431.(1946). Amer. Heart J., 31, 677.

Biorck, G. (1946a). Amer. Heart J., 32, 689.(1946b). Brit. Heart J., 8, 17.Jackson, F. S., and Rohlin, S. (1948-9). Acta med. scand., 132, 283.

Bousfield, G. (1918). Lancet, 2, 457.Boyd, L. J., and Scerf, D. (1940). Bull N. Y. med. Coll., 3, 1.Brow, G. R., and Holman, D. V. (1934). Amer. Heart J., 9, 259.Bryant, J. M., and Wood, J. E. (1947). Amer. Heart J., 34, 20.Dietrich, S, and Schweigk, K. (1933). Klin. Wchnschr., 12, 135.Evans, C., and Bourne, G. (1941). Brit. Heart J., 3, 69.Evans, W. (1952). Lancet, 2, 1092.Feil, H., and Siegel, M. L. (1928). Amer. J. med. Sc., 175, 255.Greene, O. W., and Gilbert, N. C. (1921). Amer. J. Physiol., 56, 475.__, - (1922). Amer. J. Physiol., 56, 155.Hall, D. (1932). Lancet, 1, 1254.Hecht, H. H. (1949). Arch. intern. Med., 84, 711.

92

on Novem



/B


ownloaded from


THE ANOXIA TEST FOR MYOCARDIAL ISCHAMIA 93

Katz, L. N. (1946). Electrocardiography. Kimpton, London.- , Hamburger, W. W., and Schutz, W. J. (1934). Amer. Heart J., 9, 771.Larsen, K. H. (1936). Acta med. scand., Supp., 78, 141.Levine, S. A., Ernstene, A. C., and Jacobson, B. M. (1930). Arch. intern. Med., 45, 191Levy, R. L., Barach, A. L., and Bruenn, H. G. (1938). Amer. Heart J., 15, 187.

Bruenn, H. G., and Russell, N. G. (1939). Amer. J. med. Sc.. 197, 241.and Williams, N. E. (1940). Amer. Heart J., 19, 369.

Patterson, J. E., Clark, T. W., and Bruenn, H. G. (1941). J. Amer. med. Ass., 117, 2113.Williams, N. E., Bruenn, H. G., and Carr, H. E. (1941). Amer. Heart J., 21, 634.

Lindgren, I. (1946a). Nord. Med., 30, 1114.(1946b). Nord. Med., 30, 1344.(1946c). Nord. Med., 31, 1584.

Mainzer, F., and Krause, M. (1940). Brit. Heart J., 2, 221.Malmstrom, G. (1947). Acta med. scand., Supp., 195.Mannheimer, E. (1946). J. Paediat., 29, 329.

, Carlgren, L. E., and Graf, M. L. (1948). J. Paediat., 33, 58.Master, A. M. (1935). Amer. Heart J., 10, 495.

(1953). Bull. St. Francis San., 10, 1.Friedman, R., and Dack, S. (1942). Amer. Heart J., 24, 777.Gubner, R., Dack, S., and Jaffe, H. L. (1941). Arch. intern. Med, 67, 647.Nuzie, S., Brown, R. C., and Parker, R. C. (1944). Amer. J. med. Sc., 207, 435.

Mathers, J. A. L., and Levy, R. L. (1950). Circulation, 1, 426.- (1952). Amer. Heart J., 43, 546.

Nylin, G. (1943). Nord. Med., 18, 1045.Parkinson, J., and Bedford, D. E. (1931). Lancet, 1, 15.Patterson, J. E., Clark, T. W., and Levy, R. L. (1942). Amer. Heart J., 23, 837.Penneys, R., and Thomas, C. B. (1948). Johns Hopk. Hosp. Bull., 82, 470.- , (1950). Circulation, 1, 415.Pirani, C. L., and Schlichter, J. G. (1946). Ann. intern. Med., 25, 847.Price, R. K., and Janes, L. R. (1943). Brit. Heart J., 5, 134.Pruitt, R. D., Barnes, A. R., and Essex, H. E. (1945). Amer. J. med. Sc., 210, 100.

Burchell, H. B., and Barnes, A. R. (1945). J. Amer. med. Ass., 128, 839.and Valencia, F. (1948). Amer. Heart. J, 35, 161.

Riseman, J. E. F., Waller, J. N., and Brown, M. G. (1940). Amer. Heart J., 19. 683.Roehm, D. C., Kory, R. C., and Meneely, G. R. (1952). Amer. J. Med., 13, 104.Rothschild, M. A., and Kissin, M. (1933). Amer. Heart J., 8, 729.-~, -~(1933). Amer. Heart J., 8, 745.Ruskin, A. (1947). Amer. Heart J., 34, 569.Scherf, D., and Boyd, L. J. (1948). Cardiovascular Diseases. London.Shapiro, H., and Smyth, L. (1937-8). J. Lab. clin. Med., 23, 819.Soupanki, B. (1948). Nord. med., 37, 66.Thomson, H. W., and Feil, H. (1944). Amer. J. med. Sc., 207, 588.Turner, H. B. (1933). J. Amer. med. Ass., 100, 38.Turner, R. W. D., and Morton, E. V. B. (1952). Brit. Heart J., 14, 514.Twiss, A., and Sokolow, M. (1942). Amer. Heart J., 23, 498.Weintraub, H. J., and Bishop, L. F. (1947). Ann. intern. Med., 26, 741.Wilson, F. N., and Johnston, F. D. (1941). Amer. Heart J., 22, 64.Wolferth, C. C., Bellet, S., Livezey, M. M., and Murphy, F. D. (1945). Amer. Heart J., 29, 220.Wood, F. C., Wolferth, C. C., and Livezey, M. M. (1931). Arch. intern. Med., 47, 339.Wood, P., McGregor, M., Magidson, O., and Whittaker, W. (1950). Brit. Heart J., 12, 363. on N

ovember 28, 2020 by guest. P

rotected by copyright.http://heart.bm

j.com/

Br H

eart J: first published as 10.1136/hrt.22.1.79 on 1 January 1960. Dow

nloaded from


Documents

ANOXIA TEST FOR MYOCARDIAL ISCHAEMIA · THE ANOXIA TEST FOR MYOCARDIAL ISCHAEMIA BY NORMANCOULSHED FromtheLiverpoolRegional Cardiac Centre, Sefton GeneralHospital, Liverpool ReceivedMay11,